Leonardo

Creativity in Design: Analyzing and Modeling the Creative Leap Author(s): Nigel Cross Reviewed work(s): Source: Leonardo, Vol. 30, No. 4 (1997), pp. 311-317 Published by: The MIT Press Stable URL: http://www.jstor.org/stable/1576478 . Accessed: 26/02/2013 14:24
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The MIT Press and Leonardo are collaborating with JSTOR to digitize, preserve and extend access to Leonardo.

http://www.jstor.org

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

CREATIVITY

AND

COGNITION

CONFERENCE

PAPER

Creativity Modeling

in

Design: the Creative

Analyzing
Leap

and
ABSTRACT

Nigel Cross

DESIGN CREATIVE
Significant innovations or novel design concepts are often reported as arising from sudden illuminations, or "creative Such creative leaps have for some time been regarded leaps." salto as central to the design process [1]. In some fields, the creative leap is characterized as the sudden perception of a comrepentina pletely new viewpoint on a situation. This is the basis of Koestler's model of "bi-sociation"to explain creative insight [2]. In creative design, however, a creative leap does not necof There might not essarily require a radical shift cambio viewpoint. be a transfer to a new "space",but merely a shift to a new part of the solution space and the "finding" there of an appropriate concept. This is what characterizes creative design as exploration, rather than search. Unlike bi-sociation, creative design does not necessarily consist of the making of a sudden contrary proposal, but the making of an apposite proposal. adecuada Once the proposal is made, it is seen to be an apposite response to the given, explored problem situation. It creates a match between the design requirements and the design structure of a potential new product. The sudden illumination that occurs in creative design is therefore more like building a "creative bridge" than taking a creative leap.

AN EXAMPLE OF CREATIVITY IN DESIGN

I observed an example of creative insight occurring in a design context in a protocol analysis study used at the Delft Design Protocols Workshop in 1994 [3]. The workshop was based on a set of analyses made by different researchers around the world using pre-recorded videotapes and transcripts of experimental design sessions. Two such experimental sessions were used in the Delft Workshop-one using the "think-aloud" protocols of an individual designer and the other using the naturally occurring interactions of a small team of three designers. The same hypothetical design problem was given to both the individual designer and the team: the design of a carrying/fastening device for mounting and transporting a hiker's backpack on a mountain bicycle. This device would be something like a special bicycle luggage rack. The designers were aware that the design exercise was an experiment and that they were being recorded.

A creative leap seems to have occurred as a sudden illumination in the team's design process at a point when one of the team members, Designer J, suggested the following design concept: "Maybe it's like a little vacuumformed tray." The tray idea was quickly taken up by the team, and the other members collaborated in developing the concept into a fully fledged design. The resulting design is shown in Fig. 1. The creative leap occurred at about 1 hour 18 minutes into the 2-hour design session. It was the first time a tray had been mentioned, and it seemed to provide an immediate focus for the team, whose members began to evaluate the concept in a constructive way, identifying the positive features that such a concept embodied relative to the required design features and developing the detailed aspects of the concept. The tray concept seemed to come out of nowhere after a lengthy period of exploration and problem analysis. It provided a pivotal point in the design process, after which the team focused on developing this concept into a design proposal. The team's approach to the given task was a relatively rational and systematic one. Very early in the session they planned a design strategy (Fig. 2) that was a variation of a conven-

Thesudden provided insight is widely rebya "creative leap" as a characteristic feature garded ofcreative An of design. example such a creative occurred durleap oftheactivity inga recorded study ofa small difteam. Several design ferent oftherecorded deanalyses made sign activity byindependent researchers confirm thecritical ofthiscreative role inthedeleap oftheteam. The ausignprocess thor reviews some descripgeneric tivemodels ofcreative for design into thedesign team insight study and discusses issues forthecomofcreative deputational modeling ontheseobservations. signbased Itis suggested that creative insight indesign should beregarded as a between perceptual bridge-building and rather than a solution, problem leap.
I

Fig. 1.The team's concept design.

Nigel Cross (educator), Department of Design and Innovation, Faculty of Technology, The Open University, Milton Keynes MK7 6AA, United Kingdom. E-mail: <n.g.cross@open.ac.uk>. This article is based on a paper presented at the Second International Creativity and Cognition Symposium, which was held at the Loughborough University and Loughborough College of Art and Design from 28 April to 2 May 1996.

l14 LaCt LfC< URE rMTUPtt

? 1997 ISAST

LEONARDO, Vol. 30, No. 4, pp. 311-317, 1997

311

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

. frc /*<

.4

(Rr,?e

Fig. 2. The team's design process pi-. plan.

P.740 l: ^o'43^

?Tt

F/D

^F^t .~

(l^-OCO -

One of the significant issues that arose during discussion concerned the hazard caused by the backpack's shoulriesgo der straps if they were allowed to dangle into the bicycle wheel. After generating random concept lists, the team members reviewed each list to eliminate unsatisfactory concepts and identify preferred ones. As they went through the pack-to-rack list, the "bag"concept was stressed as a solution for holding all loose straps. It was then that the tray concept suddenly appeared: I: Bag: put it in a bag. We're going to need some sort of thing to do something with thosecorrea straps. K:To get this out of the way.... J: So it's either a bag, or maybe it's like a little vacuum-formed traykind of for
it to sit in.

~ ~ f1
,s

../. . Cb^c^^

20 -5

L^^tto

tional model of the engineering design list the problem constraints; later a list process [4]. Their design process was of features they intended their product based on a model of (1) exploring the to have was added. All of these items were derived from the brief of the deproblem and writing a performance specification; (2) generating a range of sign problem and related information provided in the experiment. The team concepts; (3) evaluating and selecting the most promising concept; (4) devel- then developed the problem into three oping the concept into a detailed de- sub-problem areas: (1) the position of rack device relative to the bicycle; sign; and (5) communicating the final the estante mechanisms between (a) the (2) joining proposed design. They created a time unir schedule for completing the process, al- backpack and the rack and (b) the rack cronograma lowing about 1 hour for stages 1 and 2; and the bicycle; and (3) materials neces15 minutes each for stages 3 and 4; and sary for making the rack device. In each case, the team members exa final 30 minutes for presentation drawings and budgets. A member of the team plored problems and solutions together was appointed timekeeper/scheduler, by proposing concepts (sub-solutions) designado and he made sure that the team fol- for each sub-problem and evaluating/ lowed the plan quite closely. The cre- discussing the implications and possibilities of each concept. For example, their ative leap, which identified a concept for detailed design, occurred right on thoughts about the positioning of the ocurrido justo el objetivo rack and its supports on the bicycle target, between stages 3 and 4. Team members kept records of their frame raised issues regarding riding staworking processes in the form of bility, ergonomics of use, concerns about sketches on paper and lists compiled on the weight of a full backpack and user whiteboard. They began by attempting behavior. In general, they argued from to outline functional specifications and form to function, rather than vice versa.

I: Yeah,a tray,that'sright, O.K. J: It would be nice, I mean, just from a positioningstandpoint,if we'vegot this frameoutline and we know (backpack) that they're going to stick with that, a tray. you can vacuum-form
I: Right, or even just a small part of the

tray.... K:Somethingto dressthis (straps)in. J: Maybe the tray could have plastic snap featuresin it, so youjust like snap down into it. your backpack K:Snap in these (backpack)rails. J: It's a multi-function part. K:Youjust snap in these rails. J: Yeah, snap the rails into the tray there. I: O.K.
J: It takes care of the rooster-tail problem....

Fig. 3. Principal phases of the team's design process, identified by Giinther et al. Bars indicate focus of activity over time.

[[1113

the task clarifying

searching forconcepts

|I I 10 I 20

II

I fixingthe concept

I 30

I 40

I 50

II 60

I 70

80

II 90

I 100

I 110 120

time [min]

In this 1-minute segment, we see the key concept-the tray idea-proposed, accepted, modified, developed andjustified. As well as securely holding the backpack, a traywould solve two particular problems: the dangling straps and the "rooster-tail"problem-i.e. the water/mud spray thrown up by a mountain bicycle wheel, which creates streaks of mud on the rider's backside and which would dirty the backpack unless it were protected. The conceptual strength of the tray idea is that it embodies a type of potential solution that, once it has been expressed, recognizably satisfies certain key problems yet can be modified and refined to accommodate other problems and requirements in a satisfactory way. It is an apposite proposal. Did the tray idea come from nowhere? The above extract shows the first instance of the word "tray"in the transcript; after that the word was used repeatedly as the defining concept for the team's design proposal. (The word occurs 35 times in the last 40 min"tray" utes of the transcript.) Possibly related

312

Cross,Creativity in Design

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

concepts mentioned earlier include references to injection-molded plastic as a possible material and flat plastic forms to be used as the rack device. In fact, nearly 20 minutes before the tray idea was first expressed, its originator, Designer J, referred to a similar rack idea: "It looks like everything we're looking at right now is wire-form, but actually a friend of mine suggested a product that he would do-an injection-molded rack that would kind of like fold down-a couple of years ago...." Team member I immediately responded with a recollection of a similar device: "It's like the little rack that was flat; it had these panels ... but these panels were solid, it had little wheels . . and it would come off and then it would be like a little trailer." Designer J also suggested another kind of flat plastic panel solution a few minutes later: I think that a super simple solutionmight not be strong enough though-

120 110 . 90 80 70 o * ------Problem Solution Constraint O

Fig. 4. The team's discourse production over time (10min intervals), identified by Mazijoglou et al. Graph lines indicate changes in type of discourse.

100

Require ? """"........." Inform U rrocess

i 60
50 40
30 20 10 0o 1 2 3 4 5 6 Time 7 8 9 10 11 I I -w 12

Whiteboard

if you can imagine just taking a piece of like propylene or something like that, and die-cutting this triangle that you can fold, you know, like a cutout from a pop-up book or whatever, and it bolts on down there, and creates a flat surface ... kind of acts as a mudguard too.

Fig. 5. The team's shifting work loci over time, identified by Radcliffe. Graph points indicate things being used by the designers.

Artefact(s)

Sketches

Documents

Therefore we see that ideas related to the device as a flat sheet of plastic, which would also act as a mudguard, were suggested shortly before the concept defined as the creative leap. The significant difference in the creative leap concept is that it was the first time the proposed device was described as a tray-i.e. a flat surface with a raised lip around its circumference. (This was also the first time the process of vacuumforming was mentioned. However, as the concept was developed, the discussion of the manufacturing process reverts to injection-molding.) The tray concept summarizes a recognizably good solution in a conceivable form in a way that is significantly different from the concept of a flat, folded panel. The key difference here seems to be the perception of a tray as a container (like a bag), whereas the previous concepts had only identified a flat surface. As the earlier transcript extract showed, the tray concept seemed to be immediately recognized and accepted by the team as a good concept. However, they returned to their discipline of checking off the other concepts they had generated. But DesignerJ was careful to insist that the new concept of tray be added to the list: "I think tray is sort

20

40

60 Time (min)

80

100

120

Fig. 6. Linkograph of the team's design moves around the creative leap (move 30), identified by Goldschmidt. Letters and arrows indicate important statements by individual designers, with strong backlinks or fore-links. Numbers indicate statement in protocol section.

I> 'II

II

.>

.I1>D> .I>

<.

'^,l

-D

.,

>

I l

l<K <.

<.I

{,

(e

.,

of a new one on the list, it's not a sub-set of bag. ..." Very shortly afterwards, as they concluded this stage of their design process, Designer J also emphasized his commitment to the tray concept: "I really like that tray idea.... I think all design eventually comes down to a popularity contest." The ways in which persuasive tactics were used by the team members to get

their preferred concepts adopted, such as by expressing emotional commitment to them, have been discussed in more depth elsewhere [5]. Summarizing how the creative leap concept emerged, we can see that the concept drawsupon earlier notions that, in retrospect, seem very similar-a flat, folded surface made of plastic-but lack the apparently critical feature of containment that a tray has. The generation

Cross,Creativity in Design

313

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

COMINAlTION

MUATION

ANALOGY

PRNCIPLES

FIRSr

d
4.

m
i_'1-

v
i
"

Fig. 7. Demonstrations of the results of the procedures of combination, mutation, analogy and design from first principles, from Rosenman and Gero. In each case, a novel design results from the procedure.

i~~~i~~

~~ r~~

of the tray idea was perhaps aided by the asistida earlier consideration of a more extreme form of containment: a bag. The idea seemed to focus on one particular problem (containing the straps) as most significant, then was quickly elaborated to satisfy a range of other problems and functions. It is recognizably a bridging concept between problem and solution that synthesizes and resolves a variety of goals and constraints, and it occurred during a "review" period after earlier periods of more deliberately generating concepts and ideas.

ANALYZING THE EXAMPLE

The Delft Design Protocols Workshop was concerned with analyzing design activity across a broad spectrum of approaches; it was not concerned specifically with analyzing creativity. Of the 20 workshop papers, 10 analyzed in some form the team experiment, but none of these concentrated specifically on the creative leap identified above. However, some of the analytical modeling of design activity in these papers can be reinterpreted from the perspective of the creative leap. Most analyses of the team design process during the Delft Workshop fail to indicate how the tray concept originated, but some do reinforce the importance of this concept as marking a key point in the process. For example, Gfnther et al. [6] classified the team's protocol statements as three major stages of a design process. Their resulting chart (Fig. 3) suggests how the tray concept, which occurred at around 78 minutes, effectively ended the "searching for concepts" stage. Similarly the graph produced by Mazijoglou et al. [7] of "discourse production" (Fig. 4) shows how the team's discourse (verbal statements made) peaked in the "solution"category during the period around the emergence of the tray concept. Radcliffe's analysis [8] of the shifting "work loci" (Fig. 5) also shows how the focus shifted at around 80 minutes from handling artifacts (principally the backpack and bicycle provided for the team) and listing on the whiteboard to developing the final design largely through sketches.

The analysis that came closest to both tracing the history of the emergence of the tray concept and indicating its important role was that by Goldschmidt [9]. Her "linkograph" of the relevant section of the team protocol is shown in Fig. 6, where J's tray statement is number 30. The linkograph shows how each statement (or "move") is linked (by a "common sense" analysis of relationships between statements) to others. Statement 30 in this particular group is identified as a "critical move"-i.e. one that has a relatively high number of links to other statements that succeed it. Goldschmidt identifies this set of statements around statement 30 as a particularly "productive" phase of the team's design activity, relatively rich in interlinks between statements. Again, her analysis does not explain how the significant tray concept came to be generated, but her analysis confirms it as a statement that was influential. The linkograph (Fig. 6) shows a highly interconnected group of statements (statements 28 to 54). Statement 28 is Designer I's suggestion to "put it in a bag";statement 54 is DesignerJ's insistence that "tray is ... not a sub-set of

Fig. 8. Some emergent shapes (below) inferred from a group of three triangles (above), from Gero. The emergent shapes are discovered as implicit within the original.

bag." In that short period (2 minutes) we can see that the tray concept somehow generated a highly productive, cognitively rich sequence of interacting statements, with the team members building on each other's ideas. Statement 30 ("maybeit's like a little vacuumformed tray")does appear to come from in nowhere-it has just two "back-links" the linkograph, to the immediately preceding statements. (Other back-links-

bag = tray."

Fig. 9. Possible combinationof "panel+

a..s..
A / s \ \

I-

,/
:::::::::::::::::::::::::::::::::: :: a. .
:?? ::: ??:::::\r ??

ii
": --,

\
x

Fig. 10. Possible mutation of "flat panel" into "tray."

.'

X4
....

--

......i...i~~

Fig. 11. Possible analogy of "bag" with "tray."

A
/X

/ /:

'4?

'4~

\ x _ w

C>

314

Cross,Creativity in Design

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

for example, earlier references to flat, plastic devices-are not shown in this particular, limited section of the full session.) Its importance, however, is clear in the relatively high number of "forelinks" it has-i.e. subsequent statements that build onto, or refer back to, this statement.

Fig.2 P

l -

Fig. 12. Possible inference of design from first principles.

EXPLANATORY MODELS OF CREATIVE DESIGN

Attempts at understanding and promoting creative thinking in design generally focus on a number of techniques and procedures that utilize either free-association thinking, such as brainstorming [10], or forced associations, such as synectics [11 ]. A number of explanatory models of creative design have also been developed through research in artificial intelligence. For example, Rosenman and Gero [12] suggested four procedures by which creative design might occur: combination, mutation, analogy and design from first principles (Fig. 7). Another creative design procedure with similar potential was later added to this list by Gero [13]: emergence (Fig. 8). These procedures are widely accepted as useful explanatory models of creative design both within and outside the artificial intelligence community.

whichJ developed into "a traywith a net and a drawstring" and which K further developed into a net as something like a retractable window blind: I: Whatif your bag were big, or,whatif your, if this traywere not plastic, but like a big net, you just sort of like pulled it around and zipped there, I don't know. J: Maybeit could be a part, maybe it could be a traywith a net and a drawstringon the top of it, I like that, that's a cool idea. I: A traywith sort of just hanging down net, you can pull it around and zip it closed. K: It could be like a windowshade, so you can kind of, it sinksbackin. I: It retracts,yeah. K Youpull down,it retractsin. J: A retractingshade.

such a limit-how does a system recognize that a satisfactory or more-than-satisfactory concept has been created from combinations of previous concepts?

Mutation
Creative design by mutation involves modifying the form of some particular feature or features of an existing design. In our example, a mutation might have happened, transforming a flat panel into a tray (Fig. 10). If Designer J was thinking about the inadequacies of a flat panel (e.g. it does not securely contain the backpack), he could have thought of putting a raised lip around the edges of the panel, giving rise to the concept of a tray. Designer K's early sketch (see below, Fig. 13) also might have been influential in suggesting such a mutation. We do not know what cognitive processes gave rise to J's creative leap, but it does seem that a mutation procedure could have generated the tray idea from the flat panel idea. The difficulty for computational systems for design would be identifying which structural features of the existing design to select for modification and determining which kinds of modification to apply. In this case, to reproduce "flat panel - tray,"it would have been necessary to identify the panel edges as relevant features and modify them by thickening and/or extending them beyond the surface plane of the existing design. It would have been necessary for the

In this sequence of the team's dialogue, we can see how the initial idea of "panel + bag = tray"could develop into "bag + tray,"leading to an original concept of a tray with some form of retractable, net-bag container. The resulting Combination device might have been a more creative Creative design can occur when features combination of "panel + bag" than the from existing designs are combined into tray concept. In the end, the team did new configurations. In my example of not develop the retractable net-bag idea, the tray idea, relevant concepts in the but added cross-over straps to the tray as team's discussion preceding the creative a means of constraining the backpack. leap were those of a flat plastic panel The team seemed to know how far to and a bag. It seems possible that the crepursue novel combinations before withative leap occurred when "panel" and drawing to reconsider and start another "bag"were combined, resulting in "tray" line of reasoning. In the development of (Fig. 9). In this case, the tray was not a computational systems for design, it new kind of artifact (trays already exist), would be difficult to know how to set but the combination of "panel" and "bag"in the designer's mind could have been associated with "tray," as suggested a in Fig. 9. In terms of the team's design at that was process particular point, tray Fig. 13. Possible ina novel concept. ference of emerA more novel concept might have gent concept from arisen from the combination of "panel" previous represenand "bag," for example, a bag with a tations. flexible upper part and a rigid, flat panel bottom part (again, such artifacts do already exist). In fact, the team members did go on to propose novel combib nations of "panel" and "bag." Immediately after the initial acceptance of the tray idea, Designer I articulated the concept of a net-like zippered container,

N.A

Cross, Creativity in Design

315

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

an existing deperceived asubicadas lying within en Within the artificial sign. intelligence Fig. 14. Model of community, emergence has been disthe symmetrical recussed mainly in reference to the recoglationship between design problem and nition of emergent, or extensional, design solution, (Sub- solutions ) shapes within original, intentional (Sub- problems from Cross. shapes. These may be categorized emergent structures. However, Gero has pointed out the importance of considering functional and behavioral properties mutation procedure to be based on rec- design concepts. The example given by as well as structure in designs [18]. It ognition that a flat panel would be inad- Rosenman and Gero [15] is the design seems that designers do recognize emerof the novel "balance" chair from the equate for containment. gent behaviors and emergent functions, first principles of the ergonomics of sit- as well as emergent structures, when ting posture, such as correct support for considering solution concepts. For exAnalogy The use of analogical thinking has long the spine and buttocks. But what are the ample, DesignerJ pointed out the emerbeen regarded and suggested as a basis first principles for a carrying/fastening gent behavioral property of protection for creative design. We have already device for mounting and transporting a from rooster tail spray in the tray as a seen, in the extract above, how the idea hiker's backpack on a mountain bicycle? further benefit of the concept. What may be an attempt at design of a window shade was used to help deIn our example, it is difficult to know scribe (if not necessarily to generate) a from first principles appears in Designer whether the tray idea occurred as a case concept of a retractable net-bag. The K's sketch produced early in the team's of emergence. In this context, it is intertray idea seemed to have originated in session (shown on the left-hand side of esting that Designer K had made a close association with the bag idea. De- Fig. 12). K may have made this sketch of sketch quite early in the session (at signer J said, "So it's either a bag or "backpack + accessory + bicycle" as a per- around 40 minutes) of what could be a sonal attempt to represent the design maybe it's like a little vacuum-formed design proposal with a strong resemtray, kind of, for it to sit in," which sug- problem-she did not show the sketch blance to a tray (Fig. 13a). As with her gests that he thinks of a tray as a viable to the rest of the team, and it played no possible "first principles" sketch above, alternative to a bag. This strongly sug- overt role in the design process. How- K did not offer this second sketch to the gests the analogue "bag= tray"(Fig. 11), ever, it may express the first principles of team; she made the sketch while the based on the idea that a bag is meant to the design problem and may embody a other two team members were engaged enclose and carry something. tray-like solution concept. Designer K in another activity. However, the other The difficulty in computational mod- later sketched such a solution concept, two designers certainly became aware of the sketch later, because they both used eling based on analogy is that the appro- as discussed below. Designing from first principles is at it (at around 60 minutes) and added priate behavioral features of an existing design must be abstracted. In this ex- the core of any significant understand- some additional features to it: Designer ample, the behavioral features of enclos- ing of design-it assumes the theoretical J drew some adjustable stays onto the ing and carrying were apparently se- position that designing starts with iden- drawing, and Designer I drew the wheels lected as relevant, while other behaviors tifying requirements, or desired func- of a fold-down trailer onto it. Designer I such as flexibility were not. Further- tions, and moving from these to appro- had just previously sketched the trailer more, it would seem that partial enclo- priate forms or structures. It is the concept (Fig. 13b). sure (such as in a tray) was chosen as abductive leap of reasoning from funcOne could speculate that the tray more relevant than full enclosure (as in tion to form that is regarded as the ker- emerged as a structure from either Dea bag)-about 20 minutes before the nel of design [16]. But in practice, as signer K's sketch or from the previously tray ideaJ suggested, "Maybeit's a little was seen in the design team's protocols sketched trailer concept. However, there bucket that it sits in," but this idea was and has been suggested by others [17], is no real evidence for this. If there had ignored by the rest of the team and ap- designers start by suggesting "proto- been such evidence, the emergence proparently soon forgotten. A "bucket" is models" of forms or structures and cedure would have been one of recogmore bag-like than a tray is, but the evaluate these in order to amplify the nizing the box-like structures in the bucket analogy was apparently not requirements or desired functions. sketches and converting them to a shaldeemed to be appropriate. low box-i.e. a tray. In anything other than two-dimensional graphic or decoEmergence First Principles Emergence is the process by which new, rative design, recognizing emergence is Designing from "firstprinciples" is often previously unrecognized properties are not simply a matter of shape recogniadvocated as a way of generating good tion. It involves recognizing emergent proponer behavior out of structure and/or emerand/or creative designs [14]. First pringent function out of behavior. ciples are fundamental facts or theories Fig. 15. The duck-rabbit perceptual puzzle. that supposedly, if followed rigorously, can lead to a functional solution conDISCUSSION cept. The difficulties for both artificial and natural design processes are in Models of the design process often sugidentifying exactly what the first pringest that designing should proceed in a ciples may be in any design situation sequence of stages, such as the stage and how they may be used to generate process adopted by the team studied

Overall problem) <

( Overallsolution

316

Cross,Creativity in Design

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions

here. Such models also suggest that the overall problem should be decomposed into sub-problems, and then sub-solutions found and combined into an overall solution. This is what the team attempted. However, as was seen in this study, exploration and identification of the complex network of sub-problems is in practice often pursued by considering possible sub-solutions. In practice, creative designing seems to proceed by oscillating between subsolution and sub-problem areas, as well as by decomposing the problem and by combining sub-solutions. This practice corresponds to the explanatory model I have proposed (Fig. 14) [19]. The example in this study illustrates how creative design is manifested in the creation of an apposite concept. The appositional nature of design reasoning has been neglected in most descriptive models of the design process. During the design process, partial models of the problem and solution are constructed side by side, as it were. The crucial factor is the bridging of these two partial models by the articulation of an apposite concept (the tray idea in this example) that enables the models to be mapped onto each other. The creative leap is not so much a leap across between analysis and synthethe chasm abismo sis as it is the building of a bridge across the chasm between problem and solution. The bridge recognizably embodies satisfactory relationships between problem and solution. It is the recognition of a satisfactory bridging concept that provides the illumination of the creative flash of insight.

This recognition is a perceptual act by the designer (and by colleagues, as in this example of teamwork); and our knowledge of perceptual puzzles can provide analogies of the process. For example, the recognition of a proposed design concept as embodying both problem and solution together may be regarded as something like the wellknown duck-rabbit puzzle (Fig. 15); it is neither one nor the other, but a combination that resolves both together and allows either to be focused upon. Suggesting that "Maybeit's a little vacuumformed tray" is rather like saying, "Maybeit's a duck-rabbit." Acknowledgments
This paper is based on data from the Delft Design Protocols Workshop, 1994, organized by Kees Dorst, Henri Christiaans and Nigel Cross at Delft University of Technology, in association with Steve Harrison and Scott Minneman of Xerox Palo Alto Research Center (PARC). The workshop was made possible by the financial, practical and moral support provided by the faculty of Industrial Design Engineering of Delft University of Technology, Xerox PARCand the Engineering Design Center of Stanford University. Above all, gratitude is due to the anonymous designers who willingly participated in the experiments, provided their time and talent free of charge and allowed their design activity to be analyzed in this way.

ProductDesign (Chichester, U.K.: Wiley and Sons, 1994). 5. N. Cross and A. Clayburn Cross, "Observations of Teamwork and Social Processes in Design," Design Studies16, No. 2, 143-170 (1995). 6.J. Gunther, E. Frankenberger and P. Auer, "Investigation of Individual and Team Design Processes in Mechanical Engineering," in Cross et al. [3]. 7. M. Mazijoglou, S. Scrivener and S. Clark, "Reprein Crosset al. [3]. senting Design WorkspaceActivity," 8. D. Radcliffe, "Concurrency of Actions, Ideas and Knowledge Displays Within a Design Team," in Cross et al. [3]. 9. Goldschmidt, "The Designer as a Team of One," in Cross et al. [3]. 10. A. Osborn, Applied Imagination (New York: Scribner's, 1963). 11. W.J.Gordon, Synectics: TheDevelopment of Creative Capacity(New York: Harper, 1961). 12. M. Rosenman andJ. Gero, "Creativityin Design Using a Design Prototype Approach," in J. Gero and M.L. Maher, eds., ModelingCreativity and Knowledge-BasedCreativeDesign (Hillsdale, NJ: Lawrence Erlbaum Associates, 1993). 13.J. Gero, "Computational Models of Creative Design Processes," in T. Dartnall, ed., Artificial Intelligence and Creativity(Dordrecht, the Netherlands: Kluwer Academic Publishers, 1994). 14. M. French, Invention and Evolution: Design in Nature and Engineering (Cambridge, U.K.: Cambridge Univ. Press, 1994). 15. Rosenman and Gero [12]. 16. N. Roozenberg, "On the Pattern of Reasoning in Innovative Design," Design Studies 14, No. 1, 4-18 (1993). 17. L. March, "The Logic of Design and the Question of Value," in L. March, ed. The Architecture of Form (Cambridge, U.K.: Cambridge Univ. Press, 1976). 18. J. Gero, "Design Prototypes: A Knowledge Representation Schema for Design," AI Magazine 11, No. 4, 26-36 (1990). 19. Cross [4].

References
1. L.B. Archer, "Systematic Method for Designers," the Design Council, London, U.K. (1965). Rein DesignMethprinted in N. Cross, ed., Developments odology(Chichester, U.K.: Wiley and Sons, 1984). 2. A. Koestler, The Act of Creation (London: Hutchinson, 1964). 3. N. Cross, H. Christiaans and K. Dorst, eds., Analysing Design Activity (Chichester, U.K.: Wiley and Sons, 1996). 4. N. Cross, Engineering for Strategies Design Methods:

Cross,Creativity in Design

317

This content downloaded on Tue, 26 Feb 2013 14:24:51 PM All use subject to JSTOR Terms and Conditions